ࡱ> B D / 0 1 2 3 4 5 6 7 8 9 : ; < = > ? @ A PbjbjWW ' 55 &*WW''''''8(+ '\S7x>">>>$jx}`-//////$8"$S' dh$j SWW>>% zzz AW>'>-z -zzjS $>2PO' 2 <\Az`%`%d`%' z SSz \ `%  :NETMORPH manual NETMORPH version 2011-06-24, manual updated 2014-04-03 Randal A. Koene, Frank Postma, Sander de Ridder, Sacha Hoedemaker, Andrew Carnell, Pieter Laurens Baljon, Jaap van Pelt and Arjen van Ooyen. Department of Integrative Neurophysiology, VU University Amsterdam, The Netherlands Table of contents  TOC \o "1-3" \u 1. Introduction  PAGEREF _Toc384305741 \h 1 2. Installing NETMORPH  PAGEREF _Toc384305742 \h 2 2.1. Platform specific notes  PAGEREF _Toc384305743 \h 2 2.2. Optimization  PAGEREF _Toc384305744 \h 3 2.3. Quick start  PAGEREF _Toc384305745 \h 3 3. Running NETMORPH simulations  PAGEREF _Toc384305746 \h 3 4. Context: morphogenesis  PAGEREF _Toc384305747 \h 4 5. Prototyping protocol using model schemas  PAGEREF _Toc384305748 \h 5 5.1. Set prefix labels  PAGEREF _Toc384305749 \h 5 5.2. Schema matching by set logic  PAGEREF _Toc384305750 \h 6 5.3. Contributing identifiers  PAGEREF _Toc384305751 \h 7 5.4. Region specific populations of neurons of selected type  PAGEREF _Toc384305752 \h 7 5.5. Region specific models and parameter values  PAGEREF _Toc384305753 \h 8 5.6. How to generate pyramidal neurons with a 3D pyramidal apical dendrite  PAGEREF _Toc384305754 \h 9 6. Writing and running a simple sample simulation script  PAGEREF _Toc384305755 \h 9 6.1. Creating an editable textfile in the NETMORPH folder  PAGEREF _Toc384305756 \h 9 6.2. Adding parameters regarding size and duration of the simulation  PAGEREF _Toc384305757 \h 10 6.3. Adding growth and bifurcation parameters  PAGEREF _Toc384305758 \h 10 6.4. Adding output parameters  PAGEREF _Toc384305759 \h 10 6.5. Running the sample script  PAGEREF _Toc384305760 \h 11 6.6. Visualizing the sample script (for Cygwin users)  PAGEREF _Toc384305761 \h 11 7. A fully described sample simulation script  PAGEREF _Toc384305762 \h 12 8. USER parameters  PAGEREF _Toc384305763 \h 26 8.1. Global simulation context  PAGEREF _Toc384305764 \h 26 8.2. Neuronal populations and embedding space  PAGEREF _Toc384305765 \h 27 8.3. Morphological development: general  PAGEREF _Toc384305766 \h 28 8.4. Morphological development: growth cone bifurcation  PAGEREF _Toc384305767 \h 28 8.5. Morphological development: growth cone elongation  PAGEREF _Toc384305768 \h 30 8.6. Morphological development: growth cone direction  PAGEREF _Toc384305769 \h 33 8.7. Morphological development: neurite fiber diameter  PAGEREF _Toc384305770 \h 34 8.8. Connectivity development: synapse formation  PAGEREF _Toc384305771 \h 34 8.9. Simulation output: network data  PAGEREF _Toc384305772 \h 35 9. Advanced invocation of NETMORPH simulations  PAGEREF _Toc384305773 \h 35 10. ADVANCED parameters  PAGEREF _Toc384305774 \h 36 10.1. Global simulation context  PAGEREF _Toc384305775 \h 36 10.2. Complete simulated network  PAGEREF _Toc384305776 \h 36 10.3. Network placement constraints  PAGEREF _Toc384305777 \h 37 10.4. Morphological development: general  PAGEREF _Toc384305778 \h 39 10.5. Morphological development: growth cone elongation  PAGEREF _Toc384305779 \h 40 10.6. Morphological development: growth cone bifurcation  PAGEREF _Toc384305780 \h 43 10.7. Morphological development: growth cone direction  PAGEREF _Toc384305781 \h 44 10.8. Morphological development: neurite fiber diameter  PAGEREF _Toc384305782 \h 46 10.9. Connectivity development: synapse formation  PAGEREF _Toc384305783 \h 47 10.10. Simulation output: network data  PAGEREF _Toc384305784 \h 48 10.11. Simulation output: histological slice generation  PAGEREF _Toc384305785 \h 49 10.12. Simulation output: statistical data  PAGEREF _Toc384305786 \h 49 10.13. Simulation output: runtime options  PAGEREF _Toc384305787 \h 51 10.14. Simulation output: graphical visualization  PAGEREF _Toc384305788 \h 52 10.15. Simulation output: sequences and animation  PAGEREF _Toc384305789 \h 54 10.16. Outgrowth in a qualified environment  PAGEREF _Toc384305790 \h 55 11. Output files generated during NETMORPH simulations  PAGEREF _Toc384305791 \h 55 11.1. stats.m  PAGEREF _Toc384305792 \h 55 11.2. Textual output files containing the generated network structure  PAGEREF _Toc384305793 \h 56 11.2.1. Txt header  PAGEREF _Toc384305794 \h 56 11.2.2. Txt neurons  PAGEREF _Toc384305795 \h 57 11.2.3. Txt synapses  PAGEREF _Toc384305796 \h 57 11.2.4. Txt fiber structure root nodes  PAGEREF _Toc384305797 \h 57 11.2.5. Txt fiber continuation nodes  PAGEREF _Toc384305798 \h 57 11.2.6. Txt fiber bifurcation nodes  PAGEREF _Toc384305799 \h 58 11.2.7. Txt terminal fiber growth cones  PAGEREF _Toc384305800 \h 58 11.2.8. Txt apical dendrite tuft root nodes  PAGEREF _Toc384305801 \h 58 11.2.9. Txt apical dendrite oblique root nodes  PAGEREF _Toc384305802 \h 59 12. References  PAGEREF _Toc384305803 \h 59 13. Studies using NETMORPH  PAGEREF _Toc384305804 \h 59  Introduction NETMORPH is a simulation framework for the developmental generation of 3D large-scale neuronal networks with realistic neuron morphologies. In NETMORPH, neuronal morphogenesis is simulated from the perspective of the individual growth cone. For each growth cone in a growing axonal or dendritic tree, its actions of elongation, branching and turning are described in a stochastic, phenomenological manner. In this way, neurons with realistic axonal and dendritic morphologies, including neurite curvature, are generated. Synapses are formed as neurons grow out and axonal and dendritic branches come in close proximity of each other. NETMORPH is a flexible tool that can be applied to a wide variety of research questions regarding neuronal morphology and synaptic connectivity. Research applications include studying the complex relationship between single neuron morphology and global synaptic connectivity. The assumptions and intrinsics of NETMORPH are described in Koene et al. (2009) (Ref. [1]). NETMORPH is the result of the joint efforts of Randal Koene, Betty Tijms, Peter van Hees, Frank Postma, Sander de Ridder, Sacha Hoedemaker, Jaap van Pelt and Arjen van Ooyen of the Neuroinformatics Group at the Department of Integrative Neurophysiology, Center for Neurogenomics and Cognitive Research, VU University Amsterdam, De Boelelaan 1085, 1081 HV Amsterdam, The Netherlands. The work was supported by grants from NL NWO-CLS2003 CASPAN (635.100.005), EU MC-RTN NEURoVERS-it (019247) and EU BIO-ICT SECO (216593) awarded to Jaap van Pelt and Arjen van Ooyen. Installing NETMORPH Windows and 64-bit Linux users, please read platform specific notes below! The following steps demonstrate how to obtain and install NETMORPH. Download the NETMORPH source package. The most recent version of NETMORPH is available at: http://www.neurodynamics.nl Save the archive file netmorph-YYYYmmdd.HHMM.tar.gz in a suitable directory or folder of your choice. The version identifier YYYYmmdd.HHMM will contain version numbers of the most recent version of NETMORPH, e.g. netmorph-20090224.1225.tar.gz Extract the archive file a) Linux users may use tar, for example: tar zxvf netmorph-YYYYmmdd.HHMM.tar.gz b) Windows users may use WinRAR for Windows downloadable from  HYPERLINK "http://www.rarlab.com" http://www.rarlab.com (Windows users need to execute Cygwin Bash Shell at this point) Change the active directory to the one that was created during extraction of the archive file: cd netmorph-YYYYmmdd.HHMM; NOTE: This directory is now the main NETMORPH directory: installation, running, output storage, etc. takes place within this directory Read the files README and INSTALL to determine if any information that is relevant to installation with the newest version of NETMORPH differs from the description in this manual: less README ; less INSTALL (If the program less is unavailable, try more or another text reader.) Run the installation script within the NETMORPH directory: ./install.sh If all went well, then test NETMORPH by running it with its default settings: ./netmorph After these steps, you should have a working installation of NETMORPH and a running NETMORPH simulation that produces some characteristic results. Platform specific notes The instructions above suffice for installation of NETMORPH on 32/64-bit Linux and Mac OSX platforms, as well as for installation on Windows within Cygwin, if the Cygwin environment was ready for NETMORPH, for example due to a previous installation of an earlier version of NETMORPH. If you are installing NETMORPH in a Cygwin environment for the first time, follow the additional platform specific advice below. Windows with Cygwin: NETMORPH is tailored to a Unix operating environment. Windows users can install and use NETMORPH by providing such an environment within Windows through a collection such as Cygwin. Cygwin is free software available at http://www.cygwin.com/. Please follow the instructions there to install Cygwin, but note the points below about important programs to include during installation, so that NETMORPH will compile and will be fully functional. At some point during Cygwin installation, a list of possible download sites is presented, select one of the sites (one may choose a site from a nearby country). Subsequently a window is presented with packages that can be included in the Cygwin environment. When this opportunity is presented, choose the option to display the full list of available software by clicking the View button and select the following packages for installation in addition to the Default packages included in the Curr installation: gcc-g++ : C++ compiler make : The GNU version of the make utility Xfig : An interactive drawing tool ImageMagick : Image manipulation software suite (utilities) If Cygwin was already installed on your computer then the Cygwin Setup program is normally accessible either through a short-cut on the Windows desktop or from a Cygwin folder in the program menu. Please follow the instructions provided by the developers of Cygwin to determine the most up-to-date method of installing the necessary program packages! In Windows, a suitable directory to which to copy the archive file netmorph-YYYYmmdd.HHMM.tar.gz may be the Windows desktop or your home directory, since those normally also appear under the installed root directory of Cygwin. You can use the Unix change directory command cd to navigate to the directory containing the archive file and proceed according to the usual installation steps. To use Xfig for the visualization of generated networks in Windows, type startx at the Cygwin prompt. An X-Windows terminal should appear, and X-Windows programs such as Xfig can be run by typing their names into that terminal. Optimization Various compilers, operating systems and processor architectures allow for optimization of the program at compile time through the use of compiler flags. For reasons of maximal support, these flags have been omitted from the current makefiles. Once performance becomes an issue, it is advisable to recompile NETMORPH after setting the parameter MACHOPT. Example settings (for -march, -mtune and -mfpmath) are commented in the current Makefile. Quick start After successful compilation, it is possible to grow single-neuron morphologies or networks of realistic neurons. The program has a large number of parameters that can be set to control the individual neuron's development, as well as 'administrative issues' during simulation. These parameters are all described in the manual that can be obtained from the same location as your NETMORPH distribution. In addition, two example scripts are provided that generate the example morphologies described in Koene et al. (2009) (Ref. [1]; Figures 8B and 8D). These scripts contain appropriate settings for all parts of the simulation. The scripts can be executed by the following command in the netmorph home directory: ./netmorph include=fig8b and ./netmorph include=fig8d respectively. This results in a series of files with prefix fig8b, followed by the timestamp, and then a description of the various outputs. You can investigate the effect of changing the various parameters by looking at changes in the output statistics. The manual describes more advanced ways of inspecting the output. Running NETMORPH simulations NETMORPH is a command line program, which means that its invocation at the command prompt provides the necessary parameters for a simulation run. The complete list of sources in which NETMORPH searches for commands is given in Table 1. Importantly, the order in which sources are parsed is the table order from top to bottom, and a redeclaration of a command replaces the value of earlier declarations. Without any commands, running NETMORPH elicits a default simulation with some interesting morphogenesis results that can be used to test the program. If the current directory contains the NETMORPH program binary, type at the command prompt: ./netmorph The directory reference ./ should be included unless the directory has been added to the PATH environment variable (as defined in Linux/Unix, Mac OSX and Cygwin under MS Windows). The default simulation is specified in the resource file .nibrrc. The straightforward way to give simulation commands to NETMORPH is to append the commands at the command line. For example: ./netmorph neurons=100 days=21 Included scripts: Commands can be stored in script files, where a semicolon (;) must separate each command. Such scripts are parsed if included with an include command. Parsing is performed in depth-first manner. Given the example command line: ./netmorph include=/netmorph-defs/standard-def dt=100 include=mydirectory/myscript days=7 NETMORPH parses commands on the command line. Consequently, the following command parsing steps are taken: NETMORPH parses the file at /netmorph-defs/standard-def. NETMORPH modifies the simulation step size (dt) to 100 seconds. NETMORPH parses the file at mydirectory/myscript. NETMORPH modifies the simulated duration of morphogenesis to 7 days. Deeper nesting of commands is possible by using the include command within command line parameter (CLP) scripts. For additional advanced interface options see Chapter  REF _Ref380153352 \w 9. Table  SEQ "Table" \*Arabic 1. Sources that provide model choices and parameter values for simulations with NETMORPH. sourcedescriptioncompiled defaultsThese are the default choices and values specified in the NETMORPH executable program file.the command lineProgram input provided by instructions that follow the program name on the command line.included CLP scriptsCommand line parameters (CLP) stored in a text file and separated by semicolons (;), included by the include command. Context: morphogenesis NETMORPH has many parameters. To simplify its use, we contextualize NETMORPH. That is, we customize the user experience to a specific application context. The context described in this manual is the use of NETMORPH to simulate morphogenesis of neurons and network formation based on the spatial proximity of developing neurons. This manual is intended to facilitate getting started with NETMORPH. We assume that the NETMORPH user intends to produce simulations with information, constraints and goals similar to those of simulation results presented in Ref. [1]. NETMORPH is customized to the morphogenesis context by separating the parameter space into USER and ADVANCED parameters. Any model options and parameters that are not necessary for simulations in the context described in the preceding paragraphs of this section are considered to belong to the ADVANCED set of parameters. Explanations in this manual focus primarily on model options and parameters that belong to the USER set listed in Chapter  REF _Ref380150744 \w 8 of this manual. The ADVANCED model and parameter commands are listed in Chapter  REF _Ref380150764 \w 10 of this manual. Prototyping protocol using model schemas Before describing the model parameters, it is important to understand the so-called prototyping protocol used in NETMORPH. That is the referencing method used to apply specific models and specific parameter values to components of the simulated network. In NETMORPH each new component inherits its model choices and parameter values from its logical parent or from a prototype if there is no parent. A prototype is a declaration of a specific model choice and associated parameter values, which can schematically apply to a number of components in a simulation. Example 1: Neurite segments following a bifurcation inherit from the parent segment before the bifurcation. Example 2: The initial segment of a dendrite has no preceding parent segment and therefore inherits from the most suitable prototype. Set prefix labels A set prefix label indicates that a model choice or parameter value applies to each member of a logical set. The set is composed of an optional user-defined region specification and one of the predefined natural sets. Table 2 lists the natural sets that are predefined in NETMORPH. The full prefix format in NETMORPH is [region label.][natural set.][contributing.], where the additional contributing identifier is an arbitrary label that is used only when multiple models are chained to work together to compute expected growth results for a component of a network, as described in Chapter  REF _Ref380400910 \r \h 5.3. In the lists of commands in Chapters  REF _Ref380150804 \w 8 and  REF _Ref380150826 \w 10, wherever an asterisk * precedes a command then that command can follow a prefix with the format introduced here. Table  SEQ "Table" \*Arabic 2. Predefined "natural" sets and their prefix labels. set prefixset descriptionno prefixThe universal set, i.e. default settings.all_axons.The set of all axons, regardless of neuron type.all_dendrites.The set of all dendrites, regardless of neuron or dendrite type.all_pyramidal_axons.The set of all axons of pyramidal neurons.all_pyramidal_dendrites.The set of all dendrites of pyramidal neurons, regardless of dendrite type.all_interneuron_axons.The set of all axons of interneurons. all_interneuron_dendrites.The set of all dendrites of interneurons.all_apical_pyramidal_dendrites.The set of all apical dendrites of pyramidal neurons. Schema matching by set logic A component without a direct parent inherits from the smallest and most specific set for which a prototype has been declared. For example, using the natural sets in Table 2, a member of the set all_apical_pyramidal_dendrites can inherit model choices and parameter values from a prototype for the set all_pyramidal_dendrites. The example in Box 1 demonstrates the procedure for the growth cones of the apical dendrites of simulated pyramidal neurons in the case where an arbor elongation model is chosen and necessary parameters are set according to implicit and explicitly specified values. Beware: It is important to note that if a model choice is specified for a given set then no further schema matching is done to locate any associated parameter values that were not explicitly specified for that set. Parameters with unspecified values receive default values that were compiled into the NETMORPH program. The default values are listed in the default value column of the command lists in Chapters  REF _Ref380150970 \w 8 and  REF _Ref380150988 \w 10 of this manual. BOX 1. Explicitly specified or compiled default parameter values.  Is an arbor elongation model specified for all_apical_pyramidal_dendrites? - NO. Is an arbor elongation model specified for all_pyramidal_dendrites? - YES. Model chosen = van_Pelt. Parameters: Is the F parameter explicitly specified for all_pyramidal_dendrites? - YES. Set the F parameter as explicitly specified. Parameters: Is the nu0 parameter explicitly specified for all_pyramidal_dendrites? - NO. Use the compiled default value for the F parameter. Example script: In the above example, an arbor elongation model and its relevant parameter values are explicitly specified for the universal set. Additionally, an arbor elongation model and its relevant parameter values are explicitly specified for apical dendrites of pyramidal neurons. Any neurite arbors that are not apical dendrites of pyramidal neurons receive the universal model and universal parameter values. If the PDF model (delta) and its value (1) had not been explicitly specified for the set all_apical_pyramidal_dendrites, then the neurites of pyramidal neuron apical dendrites would use the default probability density function for arbor elongation models and corresponding default parameter values. Contributing identifiers In some cases, the simulation of aspects of neurite growth such as elongation, branching, turning or direction of growth may be best achieved if the hypotheses underlying two or more growth model can be combined in some way. To enable this, NETMORPH permits weighted chaining of many of the developmental models within each category. The following example demonstrates how chaining two different growth cone direction models can produce a desired neurite curvature as well as a general trend of directed growth to connect different simulated layers of neurons. Example script: In this example, a segment_history_tension direction model is placed at the head of a chain of direction models that is applied to members of the all_axons natural set in the universal network region (no region prefix). That direction model computes an expected direction of growth by taking into account the tensile influence of preceding neurite segment pieces up to the most recent branch point (or soma), without the possibility that the history of preceding directions is randomly truncated (dirhistory_selection=none). Specifying the label axondm indicates that another direction model should be chained into this computation of the expected direction. The reference weight of the model at the head of a chain is always 1.0, so that the second model is given equal weight by specifying a dm_weight of 1.0. The contributing direction model is a cell_attraction direction model. It computes an expected direction vector toward the centroid of a group of target neurons, for example the neurons of an adjacent region. The expected direction vectors computed by the two models after each turn in the neurite fiber are combined by the weighted sum d = dseghistension + 1.0dcellattraction. The vector d is then modified by a perturbation that applies to the combined result of the chain of computations, within a range of perturbation angles from 0 to 0.75 radians. Region specific populations of neurons of selected type When regions are used to specify the volumes in which neuron cell bodies are placed during network initialization, then it is possible to indicate specific populations of neurons that should be placed within a region, in addition to the randomized selection of a number of neurons from the general (network-wide) populations of neurons. The additional populations are specified through the value of a parameter that is formed by the combination of a region label and a neuron type identifier. It is possible to set the number of neurons in the general population to zero, so that the numbers of each type of neuron to be placed in each region are defined entirely independently.  Example script excerpt: Used in a complete script, the parameter specifications in the example above produce a network with no general population neurons, but three regions containing specific numbers of pyramidal neurons and interneurons. Region specific models and parameter values When regions are used to specify the volumes in which neuron cell bodies are placed during network initialization then it is possible to define models and parameter values that apply exclusively to the neurons that are members of a specific region. Note that regions are allowed to overlap or occupy the same space in a 3D simulation, which makes it possible to define neurons of the same type (e.g., pyramidal neurons) with different growth models and parameter values that are distributed in the same spatial volume. Using a region identifier, as well as a natural set label (see Table 2), model choices and parameter values specified are applied as a schema. Region-specific schemas are applied first, before seeking best matches among the universally applicable schemas. The network-wide universal schemas are defined as shown in the preceding examples. The following piece of an example script demonstrates model specification in a simulation with three different regions. Example script: The three layers (IV, V and VI) are composed of neuron populations with different numbers of specific neuron types. Pyramidal neurons are placed in regions IV and V, interneurons are placed in regions V and VI. Within the regions that contain interneurons, region-specific schemas define elongation rates of the axons of interneurons that differ from the elongation rates defined by network-wide schemas applied to all other neurons. How to generate pyramidal neurons with a 3D pyramidal apical dendrite A special terminal segment elongation model, pyrAD_BESTLNN, exists to aid in the prototyping of three dimensional pyramidal neurons. In situ, pyramidal neurons have a distinct shape, characterized to a large degree by a long apical dendrite with specific features: linear extension up to several hundred micrometers, oblique branches at right angles to the main trunk of the apical dendrite, and strong arborization at the end of the linear extension. When using pyrAD_BESTLNN in NETMORPH, the characteristic shape of the trunk portion of the apical dendrite is determined through the usual set of models and parameters that apply to pyramidal apical dendrites, e.g. all_pyramidal_apical_dendrites.direction_model (and its parameters), all_pyramidal_apical_dendrites.branching_model (and its parameters), etc. In this way, it is still possible to control the likelihood that the trunk of the apical dendrite bifurcates into two trunk branches, the curviness of the trunk, and more. It is even possible to target the growth of the apical dendrite at another population of neurons by including a cell attraction model in the chain of direction models (see the description of parameters and models in the following sections). The usual parameters for growth of the trunk involve very little or no bifurcation (Binf=0), very little curvature and a much greater mean elongation rate than that of basal dendrites. The pyrAD_BESTLNN model inherits its elongation function from the non-normalizing BESTL model, so that a probability distribution determines absolute elongation rates. The pyrAD_BESTLNN model is special, in that it contains parameters that describe the appearance of oblique branches to the trunk, as well as all of the models to use for growth of oblique branches and for growth (tufting) at the tip, once the apical trunk reaches its full extent. The models that can be specified uniquely for the tuft and oblique branches of an apical dendrite (identified by a unique prefex) are: terminal segment elongation models, elongation rate initialization models, terminal segment branching models, branch angle models and direction models. It is even possible to use unique arbor elongation model parameters by specifying a special terminal segment elongation model (van_Pelt_specBM). Details concerning the use of these models are included in the next section, which contains a fully described sample simulation script. Note: It is possible to create the trunk with oblique branch points as a near-initial state by setting a very large elongation rate for the apical dendrite, which results in the completion of the trunk after only a few simulation steps. Writing and running a simple sample simulation script In this example script we will only focus on getting the program running and learning how to write a script that is able to be simulated in NETMORPH. Creating an editable textfile in the NETMORPH folder For Cygwin users: - The default folder is C:\Cygwin\home\(username)\netmorph - Right Click in the folder and mouse over New and press Text Document. - Name it for instance test.txt For LINUX users: Open the editable text file using any text editor. In the first few notes it is recommended to write some information about the simulation you are currently writing. Do this by adding # (or // in C++ ) to every comment line. Example: # This is a test file for writing a simple sample simulation script Adding parameters regarding size and duration of the simulation days=15; During this simulation the neurons have been growing for 15 days. randomseed=0; By setting randomseed to 0, each time a simulation is run a different random seed will be used in the pseudo-random generator, therefore leading to different simulation results. dt=100; The simulated time interval between fixed-step updates of the developing network is set to 100 seconds. neurons=1; This parameter defines the number of neurons to grow out. It is possible to specify a virtual space in which the neurons grow; this feature is described in Chapter  REF _Ref380151057 \w 8. Adding growth and bifurcation parameters In NETMORPH it is possible to change a great number of parameters, but fortunately not every parameter has to be described in a sample script. If parameters are not used in the script NETMORPH will use its own default values for those parameters. In this example we will create one neuron using only growth and bifurcation parameters.   SHAPE \* MERGEFORMAT  * These USER and ADVANCED parameters are described in Chapters  REF _Ref380151222 \w 8 and  REF _Ref380151345 \w 10. Adding output parameters After adding all the neuron parameters, information is needed about the output of your data. For statistical analysis it is recommended to save text files of the output. Also, NETMORPH has a built-in visualizer that can be used to view the sample simulation. In this sample script only the use of the text file output is required. The other parameters are set to false or default value.  SHAPE \* MERGEFORMAT  Running the sample script - Open the Cygwin Bash Shell executable - Locate script file in the NETMORPH default folder, e.g.: cd C:/Cygwin/home/username/netmorph/ - Execute NETMORPH and include the sample script: ./netmorph include=test.txt Visualizing the sample script (for Cygwin users) NETMORPH has an analysis and visualization tool for Windows. The visualization tool is called startWinMovie.jnlp - Double-click to start this program. - Click Start Visualization! - Locate the .continuationnodes output file of the sample simulation script (NETMORPH default folder). - Click Open. - The sample neuron(s) are now viewable in the window (Fig. 1)  Fig. 1. The visual result of a neuron generated using the sample script described in Chapter  REF _Ref380151502 \w 6. A fully described sample simulation script The following is a complete sample simulation script separated into logical paragraphs that explain each portion of the script. The script defines neuronal populations that are arranged spatially in layers roughly intended to resemble the layers of the human cerebral cortex. Only the deepest layer in this simulation explicitly contains pyramidal neurons, and the script includes model parameters for the simulated growth of apical dendrites from those pyramidal neurons through successive layers. The resulting network after 21 simulated days of development is shown in Fig. 2.   Fig. 2. (top) Network generated with the full sample script after 21 simulated days of development. (bottom) Single pyramidal cell of the network. Visualization by NEURON3D. Part 1 Comments identifying the simulation script and its purpose:  Part 1 is a block of comment lines. Any line of text provided to NETMORPH in a script (or directly on the command line) that is preceded by the hash character # or by the C++ comment line convention // is considered a line of comment that is ignored by the command parser. Here, the comment block describes the version of NETMORPH with which the script is guaranteed to work, author and date stamp, and the purpose of the simulation script. Part 2 Temporal duration, temporal granularity and deterministic or non-deterministic simulation: Part 2 specifies how many days of neural development NETMORPH should simulate. By setting randomseed to 0, each time a simulation is run a different random seed will be used in the pseudo-random generator, therefore leading to different simulation results. The comment line shows that we can use a specific random seed instead to reproduce identical simulation results on every run. The simulated time interval (dt) between fixed-step updates of the developing network is set to 100 seconds. Part 3 Initializing soma morphology: In this simulation, we want our pyramidal cell morphologies to resemble those seen in cortical slices, and we want to arrange initial directions of growth of the apical dendrites to resemble the arrangement in cortical slices. For this reason, part 3 includes a number of ADVANCED parameters (Chapter  REF _Ref380151546 \w 10): The initial somatic arrangement assumes that apical dendrite grow toward the pia and that the apical dendrite is at the opposite end of the soma in relation to the axon and basal dendrites (pia_attraction_repulsion_hypothesis). Instead of randomizing the somatic locations of the root segments of basal dendrites, we specify a vector from the soma center through the center of mass of the initial dendrite roots (use_specified_basal_direction and specified_basal_direction_rel). For each pyramidal cell, the number of basal dendrites is determined by a uniform random selection between 4 and 8. The angular deviation from the center of mass vector at which dendrites can emerge is constrained between 0 and 1.5 radians. Use a model for the distribution of basal dendrite roots, which constrains placement by computing minimal angular difference such that the area between the basal minangle and maxangle is divided into equally large areas, one for each basal dendrite (pyramidal.basal.force-model=surface_division). Effectively, the basal dendrites appear to repel each other within the available area on the soma. Finally, USER parameters (Chapter  REF _Ref380152412 \w 8) in part 3 specify the ranges in which initial lengths of the root segments are selected by a uniform random function. Dendrites have initial lengths between the default minimum 9 m and default maximum 11 m, while axons have initial lengths between 22 m and 25 m. Part 4 Comments describing the model architecture:  The second comment block in part 4 of the script describes the successive cortical layers and their prevailing neurons, which are used here as a guide for the geometric placement of specific types of neurons with specific growth models. The simulation includes a number of neurons with soma in geometric regions that represent successive layers II to VI over a total distance of about 2000 m. Part 5 Defining spatial regions: Part 5 specifies that the geometric placement of soma at the onset of simulated network development (shape) is done in multiple regions. The regions are declared with the labels II, III, IV, V and VI. Part 6 Initializing region volumes and their contents: Each region that contains neuron somata has a basic disc shape. The number of pyramidal neurons and the number of interneurons are set explicitly in each region. Here, we simulate the development of pyramidal neurons only in region VI, and their interaction with developing interneurons in the other regions. By setting the total number of neurons in each region to 0 (e.g., II.neurons=0), we insure that the specified pyramidal neurons and interneurons define all neurons placed in each region. Had this number been greater than the total of pyramidal neurons and interneurons in a region then the remaining neurons in that region would receive types according to global proportions specified (see populationsize and approxproportion in Chapter  REF _Ref380151746 \w 8 and  REF _Ref380151916 \w 10). The regions specified are shown in Fig. 3.  Fig. 3. Model regions II through VI accommodating cell bodies in a cortical layer-like spatial arrangement. Regions have a thickness of 50 m, are spaced 400 m from each other and have a radius of 600 m. Part 7 Region specific, neuron type specific model parameters:  In part 7, we specify elongation model parameters that apply only to the growth cones on axons of interneurons in regions III and V. An arbor elongation model based on the phenomenological model functions published by van Pelt and Uylings (2003; Ref. [2]) governs the resources available to the elongating axon fibers. The parameter of the elongation rate is initialized to 0.0001 m/s, which is slower than the rate specified for other axons (see below). Part 8 Environment physics: In order to constrain fiber growth to the model cortical layers, and to force the tufts of rising apical dendrite fibers to spread out horizontally and fill in the volume that would be occupied by cortical layer I, we specify physical constraints of the environment, each of which receives a label (environment_physics). Here, the only label is pia. The type of environment physics to be applied is a spherical boundary (pia.physical_boundary). The sphere, which represents the confines of the cranial cavity has a radius of 5 cm, centered so that the boundary appears above region II, defining the top of the implicitly modeled cortical layer I. To achieve the desired shape of the bending fibers near the boundary, a repulsion coefficient of 0.2 (pia.spherical_c_repulse) is specified to apply within 100 m of the boundary (pia.spherical_maxrange). Part 9 Elongation model parameters, applied to the universal set: All growth cones that do not belong to a smaller set for which arbor elongation model parameters are explicitly specified use the phenomenological model functions of elongation published by van Pelt and Uylings (2003; Ref. [2]), which include competition for resources at the level of axon or dendrite arbors (arbor_elongation_model, growth_F and F_competes_with). For these growth cones, an initial elongation rate of 0.00013889 m/s is specified (growth_nu0), i.e., 12 m per day. At each growth cone, the individual elongation rate is controlled by a specific variant of the van Pelt et al. (2001; Ref. [3]) model functions (terminal_segment_elongation_model), which was designed to take into account the position of a growth cone in a dendritic tree. In prior work, that model did not alter elongation rates between bifurcation points, so that we set the perturbation parameter to a delta function with the default value of 0 (tsem.PDF). After each bifurcation, the initial elongation rates of the new branches are selected from a normal distribution with a mean value equal to the mean value of elongation rates at all other growth cones (elongation_rate_initialization_model), and are assigned so that the greater of the two new elongation rates is given to the branch with the longer initial length (length_distribution). Part 10 Branching model parameters, applied to the universal set: All growth cones that do not belong to a smaller set for which a branching model is explicitly specified use the phenomenological model functions of branching published by van Pelt and Uylings (2003; Ref. [2]) with default parameter values (branching_model). Branching and turning events are further constrained at the universal set level by enforcing a minimum node interval of 2 m. Branch angles are determined by a model of balanced forces that takes into account the elongation rates at the two growth cones after a bifurcation (branch_angle_model). Part 11 Turning model parameters: Stochastic turn events at which growing fibers change direction occur with a mean separation along the neurite fiber of 5 m. Part 12 Direction model parameters, applied to the universal set: All growth cones that do not belong to a smaller set for which a direction model is explicitly specified experience tensile influence on their direction of growth that is exerted by a weighted history of the directions, lengths and distances of preceding pieces of neurite fiber (direction_model). All the history contributes, without selecting a limited range (dirhistory_selection). An expected direction that is computed with the specified model is perturbed by a random angle up to 0.75 radians. Part 13 Elongation models parameters, applied to axons: Axon growth cones in regions other than region III and region V, which received more specific elongation model parameters in part 7 of the script, also elongate according to model functions published by van Pelt and Uylings (2003; Ref. [2]), but with a parameter of the elongation rate that is initialized to the greater rate of 0.000520833 m/s, i.e., 45 m per day. Axons in regions II, IV and VI will become significantly larger than other axons and basal dendrites over the simulated time interval. Part 14 Direction model parameters with cell attraction, applied to axons: In addition to a direction model with tensile history and default parameters, axon direction is guided by another chained-in direction model with the label axondm. The influence of that direction model is equal to the influence of the tensile model at the head of the chain of direction models, due to a relative weighting of 0.05 (all_axons.axondm.dm_weight). The contributing model (all_axons.axondm.direction_model) computes an expected direction vector according to the geometric centroid location of neurons in another region that attract the axon growth cones. By default, a region's axon growth cones are attracted to the cell somata in the next region defined. Part 15 Elongation model parameters, applied to dendrites: The elongation of growth cones of basal dendrites is governed by a non-normalizing terminal segment elongation model (all_dendrites.terminal_segment_elongation_model) that is based on the model functions published by van Pelt et al. (2001; Ref. [3]). This is a growth cone local phenomenological model that is not affected by explicit arbor-wide resource limitations, i.e. it does not request resources from an arbor elongation model. In accordance with this model, elongation rates are not perturbed between bifurcation nodes (all_dendrites.tsem.PDF=delta with value 0). The actual elongation rate of each growth cone is selected immediately after branching, in terms of absolute values (nonnorm_BESTL_length_distribution) generated according to the probability density function of the elongation rate initialization model (all_dendrites.elongation_rate_initialization_model). The mean rate chosen for this simulation is 0.0000914464 m/s, i.e., 7.9 m per day. Development models used in this simulation assume that branching is an event that takes some finite amount of time, so that we may assign to each branch an initial length selected with a normal probability density function with mean 6 m and standard deviation 5 m (all_dendrites.tsem.branch.PDF). Part 16 Branching model parameters, applied to dendrites: Growth cones of basal dendrites branch according to the phenomenological model published by van Pelt and Uylings (2003; Ref. [2]), with specific parameter values that result in a constrained number of branches and competition between the growth cones within each dendrite (all_dendrites.B_inf, all_dendrites.E and all_dendrites.E_competes_with). Most of the branching takes place during a short period of development (all_dendrites.tau). The probability of branching at individual growth cones is also influenced by the location of the growth cone, in accordance with the S parameter of the van Pelt et al. (2001; Ref. [3]) model functions (all_dendrites.TSBM and all_dendrites.S). Part 17 Command name substitution:  We substitute the short-hand APD for the set identifier all_apical_pyramidal_dendrites in all following commands. Part 18 Branching model parameters, applied to the trunk fibers of apical dendrites: Branching of the trunk of an apical dendrite is very unlikely (APD.B_inf=0.1). Part 19 Elongation model parameters, applied to the trunk fibers of apical dendrites: The local elongation model at the growth cones of apical dendrites is explicitly intended for the apical dendrites of pyramidal neurons (pyrAD_BESTLNN) and enables the specification of a probability density function for the apical trunk length (APD_tsem.trunklength.PDF with a mean of 80 m and standard deviation 2 m), as well as the transition to separate model sets to be used for oblique branches and for the tuft of branching fibers at the end of the apical dendrite trunk. We specify the label pyr1 to be used to identify model and parameter choices for the oblique branches and for the apical tuft (APD.tsem.prefix), as shown in the descriptions of parts 20 to 23 of the script. As in part 14 of the script, the growth cone local phenomenological model specified is not affected by arbor-wide resource limitations and elongation rates are not perturbed between bifurcation nodes (APD.tsem.PDF=delta with value 0). After branching into obliques or the tuft, each branch receives an initial length selected by a normal probability density function with mean 6 m and standard deviation 1 m (APD.tsem.branch.PDF). The apical dendrites, which appear at pyramidal neurons placed only in region VI of this simulation, elongate at a rapid rate so that we can see them extend through all the layers of the model cortex during the simulated time interval of neural development. To achieve this, the elongation rate intialization model (APD.elongation_rate_initialization_model) selects high rates of elongation with a mean value of 0.0010208333 m/s (i.e., 88 m/day). Part 20 Direction model parameters, applied to the trunk fibers of apical dendrites: The direction model that is applied to growth cones in the trunk of an apical dendrite (APD.direction_model) takes into account the tensile history of a fiber segment and is perturbed only by small random angles up to 0.1 radians. Part 21 Model parameters governing the placement of oblique branches on apical dendrites: A mean of 5 oblique branches appear on the trunks of apical dendrites (APD.tsem.obliques.PDF), branching at a mean angle that is orthogonal to the trunk (APD.tsem.obliqueangle.PDF). Part 22 Elongation model parameters, applied to the tuft fibers of apical dendrites: At the tuft of each apical dendrite, growth cones change direction and the rate of branching increases. To achieve this different part of the development of apical dendrites, growth cones in the tuft use a separate set of models. As in part 14 of the script, growth cones in the tuft are governed by a local phenomenological model of elongation that is not affected by explicit arbor-wide resource limitations. Elongation rates are not perturbed between bifurcation nodes (pyr1.tuft.tsem.PDF=delta with value 0). The elongation rate of each growth cone is selected in terms of absolute values (nonnorm_BESTL_length_distribution) by a probability density function of the elongation rate initialization model (pyr1.tuft.elongation_rate_initialization_model) with a mean rate of 0.0003889 m/s (about 33.6 m/day). The initial length of branches is selected by a normal probability density function with mean 6 m and standard deviation 1 m (pyr1.tuft.tsem.branch.PDF). Part 23 Branching model parameters, applied to the tuft fibers of apical dendrites: Branching in the tuft of an apical dendrite is modeled with the functions published by van Pelt and Uylings (2003; Ref. [2]), and is subject to dendrite-wide resource limitations, but does not reference the same arbor elongaiton model parameters as growth cones in the trunk of the apical dendrite. Instead, the terminal segment branching model (pyr1.tuft.TSBM) specifies the use of a specific branching model (van_Pelt_specBM) for which the parameters are obtained through commands with the pyr1.tuft prefix. In all other respects, the terminal segment branching model computes expected branching probabilities in the same way as the TSBM van_Pelt model specified in part 15 of the script, with an S parameter value of 1. The specific arbor branching model (pyr1.tuft.branching_model) results in an increased number of branches over a greater time interval (pyr1.tuft.B_inf and pyr1.tuft.tau), while there is some competition within the apical dendrite (pyr1.tuft.E and pyr1.tuft.E_competes_with). Part 24 Elongation model parameters, applied to the oblique branch fibers of apical dendrites: As in the case of the tuft, the oblique branches of an apical dendrite also develop according to a separate set of growth models. The terminal segment elongation model selected for obliques here (pyr1.oblique.terminal_segment_elongation_model) is the same as that for the tuft, while the elongation rate initialization model has a smaller mean rate of elongation of 0.0002 m/s, i.e., 17.28 m per day. Part 25 Branching model parameters, applied to the oblique branch fibers of apical dendrites: In the oblique branches of an apical dendrite, as in the tuft, the terminal segment branching model (pyr1.oblique.TSBM) specifies an S parameter of 1 and the use of a specific arbor branching model (pyr1.oblique.branching_model) with a small number of branches over an extended time interval (pyr1.oblique.B_inf and pyr1.oblique.tau) and some competition within the apical dendrite (pyr1.oblique.E and pyr1.oblique.E_competes_with). Part 26 General simulation parameters: In part 25 of the script, we specify some parameters of simulated development that apply to all neurons. Branching is not enforced at the root length of axons and dendrites that was specified in part 3 of the script (branchatinitlength). When branches occur, they can appear at any point in the piece of elongated fiber segment that was grown during the preceding development time interval (branchinsegment). Remaining neurite fiber after the branch point is redistributed to the branches and is involved in the branch angles and allocation of elongation rates as described for parts 9 and 10 of the script. Growth cones can make turns during growth between branch points (fibreswithturns) with turning and direction models as described in parts 10 and 11 of the script. Nodes at which growth cones make turns during preceding development cannot become branch points through later probabilistic selection (Abranchesatturns and Dbranchesatturns). In this simulation, we do not seek potential sites of synapses in the developing network (candidate_synapses). Part 27 Simulation runtime and sampling parameters: During the stepwise simulation of neuronal development in the network, we collect sample data at intervals of 4320 seconds (sample_dt), i.e. every 72 simulated minutes. That data is used to produce statistical output about the developing network (statsattr_collect_statistics). When development is investigated through a sequence of structural output (in part 27 of the script we opt not to do this), or is visualized as a sequence of network figures, or by animation, as described in part 29 of the script, then the sample rate also determines the intervals at which structure data is written to a text file or at which frames are produced for visualization. Part 28 Textual data parameters: A textual indicator displays progress during the simulation of network development (outattr_show_progress). In the structural data, we opt to keep track of the simulation time points at which branching and turning nodes are generated (outattr_track_nodegenesis), or at which synapses are generated (outattr_track_synaptogenesis). We also opt to produce a textual file containing the full structural data of the generated network (outattr_make_full_Txt). Part 29 Graphical visualization parameters:  In this example no animated visualization will be used. Otherwise, use the following two parts of this manual: Of the many possible visualizations of the resulting network, we choose only to produce an XFig compatible .fig file of the entire network (figattr_make_full_Fig). In the figure of the network structure, we choose to show the neuron somata (figattr_neurons), the axon fibers (figattr_presynaptic), the dendrite fibers (figattr_postsynaptic) and any synapses (figattr_synapses). We do not show abstract connections (figattr_connections) or spatial partitions created to search for candidate synaptic sites (figattr_partitions). We also choose to display a scale bar (figattr_show_scale) and axis arrows to indicate the orientation of the network visualization (figattr_show_axis). We draw the simulation time into the visualizations as a line of text (figattr_progress). Visualizations are prepared for presentation in color (figattr_use_color) with filled circles for the soma of each neuron (figattr_fill_soma). We depict terminal segments up to their growth cones instead of to the most recent branching or turning node (figattr_tsupd_visibly). Finally, we specify a color table for the elements of the visualization that differs from the default table (commands with the CT_ prefix). Part 30 Animated visualization parameters: In order to create an animated visualization of simulated network development, we opt to produce a sequence of figures at the chosen time intervals as set in part 26 of the script (figuresequence). The animation will contain an excerpt of the network environment with a specific center location and a specific distance from the center to each box wall of the excerpt volume (sequence_zoom_center and sequence_zoom_disttoedge). The specified excerpt has a volume of about 10.648 cubic millimeters. Only fibers that are spatially located within the excerpt volume are shown (figattr_fibres_nobox), and this is done regardless whether the neuron soma to which such fiber belongs lies within the volume (figattr_box_fibre_independently). Combination of the figure sequence into an animated file format is done externally (combinesequence). A geometric rotation is applied to each figure in the sequence, so that the network is rotated three times around the Z-axis (autorotatesequence and ROT_interval_) from a specified initial angle of rotation around the Y-axis (ROT_). The resulting animation is available at: http://netmorph.org/. USER parameters Change these parameters to modify the results of simulated morphogenesis. Table rows with a yellow background signify commands that choose models. The rows that follow them list associated model parameters. Global simulation context parameter labelunits and formatdefault valuedescriptionreferenceseconds(s), decimal(see days)Morphogenesis is simulated from 0.0 to this number of seconds. Note: Use only the seconds or days command. Seconds takes precedence over days.t in Ref. [1]days(days), decimal21Morphogenesis is simulated from 0.0 to this number of days. Note: Use only the seconds or days command. Seconds takes precedence over days.t in Ref. [1]dt(s), decimal100Fixed time step size between successive model calculations.t in Ref. [1]randomseedinteger0The "master" value used to seed pseudo random generators. Reuse a randomseed value to reproduce identical simulation results. The special value 0 implies that the system clock provides the seed value, producing essentially unpredictable simulation results.includefile pathnoneInclude commands stored in the specified file.Chapter  REF _Ref380152089 \w \* MERGEFORMAT 2.3 of this manual. Example script (as in .morphogenesis.user.clp): Neuronal populations and embedding space parameter labelunits and formatdefault valueDescription neuronsinteger9The total number of neurons in the simulation.populationsizeintegerThe number of neurons of a specified type (see labels listed below table). Note: See for default value the description of the way in which NETMORPH determines population sizes below this table.*L0(m) decimal range9, 11Range of the minimum and maximum length of initial dendrite and axon segments at the onset of simulation.shapeshape nameregionsSpatial constraints for the placement of somata. Here we describe only the specification of the regions shape and its parameters. For more options see Chapter  REF _Ref380152446 \w \* MERGEFORMAT 10.regionsspace separated listpyrlayerA list of region labels, each separated from the next by a space character.shapeshape namediscSpatial constraints for the placement of somata within . Shape names are: disc, box, sphere.neuronsinteger0The number of neurons from the general pool to be placed in the .minneuronseparation(m), decimal75A placement constraint specifying the minimum distance between the centers of neighboring somata.center(m), decimal(0,0,0)The center coordinates of the .[region.]shape.radius(m), decimal700Radius of a disc or sphere shaped region.[region.]shape.thickness(m), decimal500Thickness of a disc shaped region. Note that this parameter also allows a disc to have the appearance of a cylinder.[region.]shape.(m), decimal(see above)Extent (length) of the box region shape in the dimension (see labels listed below table).: pyramidal, interneuron, bipolar, multipolar_nonpyramidal, principal_neuron, untyped_neuron ( Neuron base classes that are not intended to be used in simulations.) : width (i.e. x-axis), height (i.e. y-axis), depth (i.e. z-axis). How NETMORPH determines population sizes: Three types of commands control the population numbers of each kind of neuron in a simulation: neurons, populationsize and approxproportion (see Chapter  REF _Ref380152480 \w 10). The total number of neurons is used in conjunction with the approxproportion commands and is ignored when populationsize commands are given. When at least one populationsize command is provided with a number greater than 0, then absolute population sizes are assumed to be specified. In that case, the sizes of populations not specified are assumed to be 0 and the total number of neurons in the simulation is derived from the sum of the absolute population sizes. Otherwise, the total number of neurons is given by the neurons command or its default value 9. Allocation of those neurons to populations of neuron types is then done according to approximate proportions, which can be specified in approxproportion commands, as ratios that should add up to 1.0. Ratios not specified use default values. If nothing is specified, then the default allocation of 70% pyramidal neurons and 30% interneurons is used. Example script:  Morphological development: general parameter labelunits and formatdefault valuedescriptionfibreswithturnsflagtrueThis flag selects whether turning models are applied to the dendrite and axon fibres throughout a simulation. The choice has a significant effect on computational resource requirements. Morphological development: growth cone bifurcation parameter labelunits and format default valuedescription *branching_modelmodel labelvan_PeltThe model used to determine the probability that branching occurred at any of the growth cones of an axon/dendrite arbor during the most recent simulation time interval. The following rows describe the parameters used by the van_Pelt model. For more options see Chapter  REF _Ref380152507 \w \* MERGEFORMAT 10.*B_infdecimal4.75The asymptotic value of the expected number of branching events at an isolated segment.*tau(s), decimal319680The exponential time coefficient of the function for the expected number of branching events at an isolated segment.*Edecimal0.5The parameter governing competition between growth cones for branching events.*E_competes_withcompetition labelwhole_neuronThe number of growth cones that are included in competition if E>0. Possible choices are: whole_neuron, all_axons, all_dendrites, same_arbor.*TSBMmodel labelvan_PeltThe model used to determine the branching probability for a specific growth cone, which may involve local data, such as the centrifugal order of a growth cone or environmental influences in the vicinity of a growth cone. The following rows describe the parameters used by the van_Pelt and van_Pelt_specBM models. The van_Pelt_specBM terminal segment branching model behaves like the van_Pelt TSBM, but is intended to be specified as the TSBM of pyramidal apical dendrite tuft and oblique branches. While this TSBM also expects branching to be governed by arbor-wide constraints, it references a separate arbor branching model for which the model selection and parameter values are expected to be specified, preceded by .tuft. or by .oblique. labels. See also parts 22 & 24 of the example in Chapter  REF _Ref380152211 \w \* MERGEFORMAT 7 of this manual.*Sdecimal0The parameter for the dependence of the branching probability on the centrifugal order of a specific growth cone. (This parameter is used by the van_Pelt and van_Pelt_specBM terminal segment branching models.)*branch_angle_modelmodel labelBalanced_ForcesThe model used to determine the branch angles when a bifurcation occurs at a specific growth cone. The following rows describe the parameters used by the Balanced_Forces model. For more options see Chapter  REF _Ref380152600 \w \* MERGEFORMAT 10.*bam.bfbam.PDFPDF labelnormal (mean = /2, std=0.5, trunc = -0.1)A parameter of the Balanced_Forces model. A PDF that is used to determine the angle in the parallelogram of forces between daughter branches. This therefore also determines the angle between daughter branches, prior to perturbation according to *bam.PDF.*bam.PDFPDF labelnormal (mean=0, std=0.3, trunc=1)Probability density function used for the random perturbation of the calculated balanced forces angles. Note: A delta PDF removes perturbation.(Ref. [1]) Note about distributing the remainder of elongated fiber length after a bifurcation: To determine the random values X1 and X2 for the branch ratio, only a uniform PDF is currently implemented. When branching occurs, the actual branching location can be anywhere along the elongated fiber that was created during the last simulation interval. Fiber that protrudes beyond the determined branching location is redistributed to the new daughter branches. The ratios X1/(X1+X2) and X2/(X1+X2) determine the respective proportions of the remaining fiber given to daughter branch 1 and daughter branch 2. Morphological development: growth cone elongation parameter labelunits and formatdefault valueDescription *arbor_elongation_modelmodel labelvan_Pelt (for universal, all_axons, all_apical_pyramidal_dendrites)The model used to determine total elongation resources available to an axon/dendrite arbor at each time point. The following rows describe the parameters used by the van_Pelt model. Note: The arbor elongation model and all its parameters are ignored by growth cones that use the nonnorm_BESTL or pyrAD_BESTLNN terminal segment elongation models. For more options see Chapter  REF _Ref380152623 \w \* MERGEFORMAT 10.*growth_Fdecimal0.39 (universal) 0.16 (all_axons) 0.5 (all_apical_pyramidal_dendrites)Elongation competition factor that determines the effect of the number of branches on the rate of elongation at each terminal segment. F in Ref. [1]*growth_nu0(m/s), decimal0.00013889 (universal) 0.0005208333 (all_axons) 0.00013889 (all_apical_pyramidal_dendrites)Elongation speed/rate. If F=0, then this is the rate applied to each terminal segment. If F=1, then it is the rate for the elongation of a whole dendrite/axon arbor. 0 in Ref. [1]*F_competes_withcompetition labelsame_arbor (universal, all_axons, all_apical_pyramidal_dendrites)The number of terminal segments that are counted for competition if F>0. Possible choices are: whole_neuron, all_axons, all_dendrites, same_arbor. n(t) in Ref. [1]*aem.PDFPDF labelnormal (universal, all_axons, all_apical_pyramidal_dendrites)Probability density function used for the random perturbation of the computed arbor elongation. PDF labels and PDF specific parameters are described below this table. Note: A delta PDF removes perturbation.*terminal_segment_elongation_modelmodel labelBESTL (universal)The model used to determine the ratio of elongation resources allocated to specific terminal segments at each time point. The following rows describe the parameters used by the BESTL, nonnorm_BESTL and pyrAD_BESTLNN models. The nonnorm_BESTL model does not apply normalization to elongation speeds with regard to other growth cones. Note: Growth cones that use the nonnorm_BESTL or pyrAD_BESTLNN models ignore arbor elongation models and their parameters, in effect implying independent elongation (*growth_F=0)! The pyrAD_BESTLNN model behaves like the nonnorm_BESTL model, but is explicitly designed to enable the simulation of pyramidal apical dendrite development. For more options see Chapter  REF _Ref380152653 \w \* MERGEFORMAT 10. *tsem.branch.PDF(m), PDF labelnormal (mean=2.0, std=1.0)Probability density function used for the random selection of initial length at BESTL branches.*tsem.PDFPDF labeldelta (universal)Probability density function used for the random perturbation of the expected quota. Note: A delta PDF removes perturbation. Beware: The perturbation is added to 1.0 before multiplication with the calculated elongation speed, so that a delta PDF with a value greater than zero can cause a ramp up to infinite speed with unpredictable results!*tsem.trunklength.PDF(m), PDF labelnormal (mean=700.0, std=100.0)Probability density function used to select the trunk length for an apical dendrite in terms of the expected distance from the soma center at which the apical dendrite begins to tuft. (A parameter used by the pyrAD_BESTLNN model)*tsem.obliques.PDFPDF labelnormal (mean=7, std=3)Probability density function used to select the number of oblique branches of an apical dendrite. (A parameter used by the pyrAD_BESTLNN model)*tsem.prefixprefix labelpyrAPA label that is used as a prefix to identify models and parameters specified for use in the tuft or oblique branches of an apical dendrite. (A parameter used by the pyrAD_BESTLNN model).oblique.miscellaneousmodels and parameter values of the universal network region all_pyramidal_dendrites setThe prefix specified by *tsem.prefix and the .oblique. keyword precede a collection of model and parameter specifications for use in oblique branches of an apical dendrite. The models for which model choices and parameter values can be specified in this manner are: terminal segment elongation model, elongation rate initialization model, terminal segment branching model, branch angle model, direction model. (A parameter used by the pyrAD_BESTLNN model. Also note the van_Pelt_specBM model described as an option for *TSBM.) See parts 23 & 24 of the example in Chapter  REF _Ref380152247 \w \* MERGEFORMAT 7 of this manual..tuft.miscellaneousmodels and parameter values of the universal network region all_pyramidal_dendrites setThe prefix specified by *tsem.prefix and the .tuft. keyword precede a collection of model and parameter specifications for use in the tuft of an apical dendrite. The models for which model choices and parameter values can be specified in this manner are: terminal segment elongation model, elongation rate initialization model, terminal segment branching model, branch angle model, direction model. (A parameter used by the pyrAD_BESTLNN model. Also note the van_Pelt_specBM model described as an option for *TSBM.) See parts 21 & 22 of the example in Chapter  REF _Ref380152278 \w \* MERGEFORMAT 7 of this manual.*elongation_rate_initialization_modelmodel labellength_distribution (universal)The model used to determine the initial ratio of elongation resources expected after a bifurcation. The following rows describe the parameters used by the length_distribution model. For more options see Chapter  REF _Ref380152675 \w \* MERGEFORMAT 10. *eri.PDFPDF labelnormal (mean=0, std=1.0, trunc=3.0)Probability density function used to draw random numbers for new relative elongation rates of the two daughter branches. Note: A delta PDF leads to identical ratios.(Ref. [1]) PDF modifying parameters: delta: